Paleo-European languages

[Ares Land]

Sigh. None of those three is widely accepted.

I think that's Tavi's point.

He mentioned "PIE" and "comparative linguistics" in general. Comparative linguistics isn't just about handling a large number of words any more than medicine is about handling a large numbers of body parts.

Unfortunately, we're doing history, not physics. Sometimes scientific method can be used to explore the credibility of the story, but that seems rare. What happens seems more akin to natural history than to science.

Can you give me an example of what would be a falsifiable prediction?

Yes, Grimm's law, which I mentioned just above. You make the hypothesis that Germanic *f will match with *p in Sanskrit, Greek, or Latin.
That hypothesis is falsifiable, it was tested, and indeed we found a long list of matches with similar meanings. And, ultimately, it was falsified and amended with Werner's law.
Even the principle of regular correspondances could be tested and falsified on related languages: we find correspondances between French and Italian, fewer between French and German, none between French and (pre-Colonial!) Nahuatl.

I mean, it all relies on falsifiable predictions.

Again, you brought up *kombero, how do we know the resemblance isn't just coincidence?

I think we can't. Now, with a set of words, we can start to say that it is highly implausible that it is all coincidence, but we can't necessarily say which ones are coincidence.

We can calcute the probability of random coincidence, see here: https://www.zompist.com/chance.htm

I'll just quote a table to get an order of magnitude:

Code: Select all

Form 	#C 	#V 	Lexicon size 	Phonetic leeway	Expected matches
CV 		20 	10 	200 			2 				40
CV 		20 	10 	200 			3 				90
CV 		20 	10 	200 			5 				250
CVC 	20 	10 	4000 			2 				80
CVC 	20 	10 	4000 			3 				270
CVC 	20 	10 	4000 			5 				1250
CVCV 	20 	10 	40000 			2 				160
CVCV 	20 	10 	40000 			3 				810
CVCV 	20 	10 	40000 			5 				6250 

Look at those figures under 'Expected matches'...
So far we have, what, three "Baltic cognates"? We're practically certain to find a good deal more chance resemblances between Gaulish and Baltic.
Coromines quotes what, a dozen? Still well under the number of chance resemblances we'd expect to find.

If we had, I don't know, something like a hundred Gaulish-Baltic cognates, we could take that idea seriously.
So far, what we have is :
an Egyptian ~ Basque ~ Sanskrit correspondance?
a Gaulish ~ Lepontic ~ Sanskrit correspondance?
a Gaulish word (unattested, I should add and merely inferred on the basis of French) ~ Baltic
another Gaulish word, again, unattested and merely inferred to exist on the basis of French ~ Baltic.

For two of the words I checked, the accepted etymology isn't even the one commonly accepted.
So unless I accept that Tsalkubilos has supernatural insight into the meaning of unattested Gaulish words, I'd say the semantic link is extremely spurious...

So, given the evidence offered, I'm entirely certain there's nothing to Tsalkubilos' claims whatsoever.

[Moose-tache]

This thread in a nutshell. This business of substrates, satem Wanderwörter, secret IE sister languages, and pan-Mediterranean language families is all pure speculation. It's a human desire to see patterns in everything let loose on a stack of dictionaries.

Comparative linguistics is speculative per se, and this includes the beloved Proto-Indo-European (PIE). ... And the thing is large sets of data ("cognates") create the illusion you're doing "science".

Personally I think there is a big difference between saying "PIE probably existed because of these several thousand cognates explainable through completely regular sound changes" and "Sardinia has to be the land of the Shrdn because they both have similar consonants." The amount of errant speculation in mainstream PIE theories is minimal. It's not zero, but the fact that it's not zero does not mean that what you're doing is on a level with the reconstruction of Proto-Germanic or something. Your isolated, unsystematic, statistically empty reconstructions do not become better simply because real scientists are only mostly following proper methodology, and pointing out flaws in other linguistics projects does nothing to address the inherent flaw of assigning meaningful patterns to random noise.

[Talskubilos]

a Gaulish ~ Lepontic ~ Sanskrit correspondance?

Cisalpine Gaulish, to be precise.

a Gaulish word (unattested, I should add and merely inferred on the basis of French) ~ Baltic
another Gaulish word, again, unattested and merely inferred to exist on the basis of French ~ Baltic.

For the Gaulish words, I'd refer you to the existing bibliography, namely Xavier Delamarre's Dictionnaire de la langue gauloise.

For two of the words I checked, the accepted etymology isn't even the one commonly accepted.

???

So unless I accept that Tsalkubilos has supernatural insight into the meaning of unattested Gaulish words, I'd say the semantic link is extremely spurious...So, given the evidence offered, I'm entirely certain there's nothing to Tsalkubilos' claims whatsoever.

I've got a good intuition, but never mind. BTW, my nickname is Talskubilos, actually an Iberian anthroponym.

[Talskubilos]

Personally I think there is a big difference between saying "PIE probably existed because of these several thousand cognates explainable through completely regular sound changes" and "Sardinia has to be the land of the Shrdn because they both have similar consonants." The amount of errant speculation in mainstream PIE theories is minimal. It's not zero, but the fact that it's not zero does not mean that what you're doing is on a level with the reconstruction of Proto-Germanic or something.

I'd remind you proto-languages aren't real entities but theoretical constructs, and in the case of PIE a huge oversimplification, I'm afraid.

Your isolated, unsystematic, statistically empty reconstructions do not become better simply because real scientists are only mostly following proper methodology, and pointing out flaws in other linguistics projects does nothing to address the inherent flaw of assigning meaningful patterns to random noise.

I agree with "isolated" but I disagree with everything else.

[Ares Land]

an Egyptian ~ Basque ~ Sanskrit correspondance?

I didn't contemplated Egyptian here.

The discussion started with a š3rdn - śárdha- connection, and at the end of the day š3rdn is from Egyptian sources.

a Gaulish ~ Lepontic ~ Sanskrit correspondance?

Cisalpine Gaulish, to be precise.

Which doesn't make a difference.

For the Gaulish words, I'd refer you to the existing bibliography, namely Xavier Delamarre's Dictionnaire de la langue gauloise.

Neither *borua nor *kombero are directly attested, nonetheless...

For two of the words I checked, the accepted etymology isn't even the one commonly accepted.

???

We've been through *kombero before.
bourbe is usually listed as from *borwa 'hot spring'.

So unless I accept that Tsalkubilos has supernatural insight into the meaning of unattested Gaulish words, I'd say the semantic link is extremely spurious...So, given the evidence offered, I'm entirely certain there's nothing to Tsalkubilos' claims whatsoever.

I've got a good intuition, but never mind. BTW, my nickname is Talskubilos, actually an Iberian anthroponym.

Ah, sorry about that, Talskubilos.

I'd remind you proto-languages aren't real entities but theoretical constructs, and in the case of PIE a huge oversimplification, I'm afraid.

Yes, they are. So what?
What actual ancestors of Celtic, Baltic, Slavic, Indic, etc. speakers spoke is one thing, it's irretrievable and PIE is certainly a huge oversimplification of it.
But the important bit isn't the reconstruction, which is a useful construct, but the amply demonstrated fact that these language are related.

On the subject of your Satem-Baltic connection, you have, so far, demonstrated nothing.
I don't doubt that your intuition is top notch, but you need to substantiate it with facts.

[Talskubilos]

The discussion started with a š3rdn - śárdha- connection, and at the end of the day š3rdn is from Egyptian sources.

OK. I've corrected that.

We've been through *kombero before.

I'm afraid this is a quasi-homophony between Gaulish and a genuine Celtic word.

bourbe is usually listed as from *borwa 'hot spring'.

Which is a complete semantic mismatch.

Neither *borua nor *kombero are directly attested, nonetheless...

Please see Delamarre. Nevertheless, *komboro- is Meyer-Lübke's (REW) and Coromines's reconstruction.

On the subject of your Satem-Baltic connection, you have, so far, demonstrated nothing. I don't doubt that your intuition is top notch, but you need to substantiate it with facts.

I've already presented you them. The pity is you don't accept them.

[Ares Land]

We've been through *kombero before.

I'm afraid this is a quasi-homophony between Gaulish and a genuine Celtic word.

For the umpteenth time, why is that? what proof do we have?

bourbe is usually listed as from *borwa 'hot spring'.

Which is a complete semantic mismatch.

I agree that it's a stretch, but so is a connection with Baltic.

Neither *borua nor *kombero are directly attested, nonetheless...

Please see Delamarre.

Okay, let me see:
Le français standard bourbe vient d'un mot gaulois *bortua
désignant probablement au départ une source boueuse, DELF 79, DHLF 262.
Le mot du latin médiéval combrus qui signifie 'abattis d'arbres' (6e-7 e s.) et
'barrage' (lIe s.) ainsi que le vieux-français combre 'barrage sur une rivière' remontent
à un gaulois *com-beros, construit comme des mots du celtique insulaire v.irl. commar,
gall. cymer, bret. kemper (> Quimper) 'confluent' (*com-beros) [...]
Comberanea, IEW 132 .
Do you know what that star means? It means 'unattested'!

For that matter, why is Delamarre to be trusted when he infers Gaulish words from French sources, but not when he lists Celtic etymologies?

I should mention that he gives 'combre' as 'a dam on a river' not as 'heap'. Neither combre nor *comberos have anything to do with hills (your proposed Baltic connection)!

On the subject of your Satem-Baltic connection, you have, so far, demonstrated nothing. I don't doubt that your intuition is top notch, but you need to substantiate it with facts.

I've already presented you them. The pity is you don't accept them.

Come on. You have presented four words, each from a different place. With so few words, random chance is extremely likely.

But, you know, as far as I'm concerned, this discussion is over. We're just running around in circles. Since you desperately want that Baltic connection, so be it.

[Talskubilos]

We've been through *kombero before.

I'm afraid this is a quasi-homophony between Gaulish and a genuine Celtic word.

For the umpteenth time, why is that? what proof do we have?

There're two quasi-homonymous words: Celtic *kom-bero- 'confluence (of rivers)' and Gaulish *komboro- 'heap, accumulation'.

bourbe is usually listed as from *borwa 'hot spring'.

Which is a complete semantic mismatch.

I agree that it's a stretch, but so is a connection with Baltic.

Really? It's rather on the contrary: 'sludge' ~ 'dirt' ~ 'morass' ~ 'dregs', etc.

Do you know what that star means? It means 'unattested'!

Not exactly. It means reconstructed, just as Proto-whatever languages. The thing is many (if not most) Gaulish words aren't directly attested in known texts but they've got descendants in Romance.

For that matter, why is Delamarre to be trusted when he infers Gaulish words from French sources, but not when he lists Celtic etymologies?

Because he's an academic guy and so he follows the established theories.

I should mention that he gives 'combre' as 'a dam on a river' not as 'heap'. Neither combre nor *comberos have anything to do with hills (your proposed Baltic connection)!

As I said before, the meaning 'heap, accumulation' is Coromines' reconstruction and matches nicely Basque (Biscayan) gonburu, bonburu 'excess', usually derived from Latin cumulu-.

Come on. You have presented four words, each from a different place. With so few words, random chance is extremely likely.

Ha, ha.

[Talskubilos]

Since you desperately want that Baltic connection, so be it.

I could add more material from Etruscan/Latin. For example, Latin caesius 'light blue' would be a loanword from Etruscan caisie-, ceisi(e)-, seemingly related to caiz-na, ceiz-na, which in turn would correspond to Lithuanian gaĩsa-s ‘glow, redness in the sky’, Latvian gàiss ‘air, wheather’, gàišs ‘bright, clear’, gaisma ‘light’, with Etruscan c ~ Baltic g, Etruscan z ~ Baltic ss and Etruscan n ~ Baltic m.

[Linguoboy]

Since you desperately want that Baltic connection, so be it.

I could add more material from Etruscan/Latin. For example, Latin caesius 'light blue' would be a loanword from Etruscan caisie-, ceisi(e)-, seemingly related to caiz-na, ceiz-na, which in turn would correspond to Lithuanian gaĩsa-s ‘glow, redness in the sky’, Latvian gàiss ‘air, wheather’, gàišs ‘bright, clear’, gaisma ‘light’, with Etruscan c ~ Baltic g, Etruscan z ~ Baltic ss and Etruscan n ~ Baltic m.

So, just so I understand: "mire" ~ "hot springs" is too much of a semantic mismatch, but not "redness in the sky" ~ "light blue"?

[Talskubilos]

So, just so I understand: "mire" ~ "hot springs" is too much of a semantic mismatch, but not "redness in the sky" ~ "light blue"?

Different colors of the sky, but never mind.

Due to the deep influence of Etruscan aristocracy in ancient Rome, Latin has apparently a lot of Etruscan loanwords. For example, just as the personal name Tiberius corrresponds to Etruscan Thefarie, the hydronym Tiber would have originated from *thefari 'river' (cfr. Hurrian tem-ari 'irrigation ditch, canal'), later thepri, thefri.

[Linguoboy]

So, just so I understand: "mire" ~ "hot springs" is too much of a semantic mismatch, but not "redness in the sky" ~ "light blue"?

Different colors of the sky, but never mind.

The sky can also be golden, greenish, grey, violet, black, etc. Very handing for finding cognates among colour terms!

Due to the deep influence of Etruscan aristocracy in ancient Rome, Latin has apparently a lot of Etruscan loanwords. For example, just as the personal name Tiberius corrresponds to Etruscan Thefarie, the hydronym Tiber would have originated from *thefari 'river' (cfr. Hurrian tem-ari 'irrigation ditch, canal'), later thepri, thefri.

Unless of course the borrowing goes the other way and Tiber actually has an Indo-European etymology.

[Talskubilos]

Due to the deep influence of Etruscan aristocracy in ancient Rome, Latin has apparently a lot of Etruscan loanwords. For example, just as the personal name Tiberius corrresponds to Etruscan Thefarie, the hydronym Tiber would have originated from *thefari 'river' (cfr. Hurrian tem-ari 'irrigation ditch, canal'), later thepri, thefri.

Unless of course the borrowing goes the other way and Tiber actually has an Indo-European etymology.

Which one? Surely our friend Ares Land would say it's a coincidence.

[Travis B.]

Due to the deep influence of Etruscan aristocracy in ancient Rome, Latin has apparently a lot of Etruscan loanwords. For example, just as the personal name Tiberius corrresponds to Etruscan Thefarie, the hydronym Tiber would have originated from *thefari 'river' (cfr. Hurrian tem-ari 'irrigation ditch, canal'), later thepri, thefri.

Unless of course the borrowing goes the other way and Tiber actually has an Indo-European etymology.

Which one? Surely our friend Ares Land would say it's a coincidence.

And who says words must be borrowed into IE; what evidence do we have these words aren't IE loanwords in Etruscan?

[Talskubilos]

And who says words must be borrowed into IE; what evidence do we have these words aren't IE loanwords in Etruscan?

in that case, we should find a suitable IE etymology.

[Richard W]

And who says words must be borrowed into IE; what evidence do we have these words aren't IE loanwords in Etruscan?

in that case, we should find a suitable IE etymology.

I'd start the 'PIE' bidding as an Umbrian name. Should an Umbrian river name have an Umbrian etymology? I ask because English river names usually don't have English (folk) etymologies.

[Richard W]

Unfortunately, we're doing history, not physics. Sometimes scientific method can be used to explore the credibility of the story, but that seems rare. What happens seems more akin to natural history than to science.

Can you give me an example of what would be a falsifiable prediction?

Yes, Grimm's law, which I mentioned just above. You make the hypothesis that Germanic *f will match with *p in Sanskrit, Greek, or Latin.
That hypothesis is falsifiable, it was tested, and indeed we found a long list of matches with similar meanings. And, ultimately, it was falsified and amended with Werner's law.

I'm surprised at how well that prediction works. Looking at words surviving from Germanic into modern English, I found seven exceptions. The best are four (mismatch), five (match-mismatch, unless either *f matching is good enough) and fork (borrowing into Germanic). The other four are flay, flock, flow and folk - no cognates.

[KathTheDragon]

The standard explanation for "four" is lexical analogy to "five", something known to be extremely common (i.e. numbers being influenced by adjacent or near-adjacent numbers)

[Talskubilos]

I'd start the 'PIE' bidding as an Umbrian name. Should an Umbrian river name have an Umbrian etymology? I ask because English river names usually don't have English (folk) etymologies.

I still think an etymology from Etruscan 'river' would be better. Anyway, this thread is about Paleo-European languages.

[Richard W]

The standard explanation for "four" is lexical analogy to "five", something known to be extremely common (i.e. numbers being influenced by adjacent or near-adjacent numbers)

So how do you rephrase the prediction? There's an alternative derivation from initial *pʷ, which improves the 'match' with Afroasiatic. I was asking for a falsifiable prediction. The concept of *pʷ has the problem of irregularly relaxing to *kʷ - we seem to do just as well positing irregular changes of *kʷ to *p. Do we also exclude 'five' from the test set as being a number?

The word wolf arguably satisfies the prediction because of Latin lupus - unless we rule that out as a Sabine loanword and say Latin has /p/ but does not exhibit *p.